Re: The spread of pinching rollers this year
 

I agree with Ike, thanks for keeping us honest!

Re: The spread of pinching rollers this year
 

Hi Ike,

Actually, our roller runs at over 20 ft/sec so that when we're driving full bore at 10 FPS we can grab a ball on the fly with no "bounce-back". If we reverse it at that speed it actually is quite a nice kicker! In fact, our autonomous simply does that - reverses the roller to full speed, we drive about 2 feet into the ball in front of us and it spins the ball up to a high enough speed to run it right up the ramp into the goal!

John

Re: The spread of pinching rollers this year
 

Wow, an that is why it was my favorite collector this year.

Re: The spread of pinching rollers this year
 

Can somebody explain the physics behind the "intake must run at >=2x robot speed" assertion? I've heard this before, and even designed with it in mind, but only now do I stop to think about it.

Intuitively, I don't see why running at >1x robot speed wouldn't work?

Re: The spread of pinching rollers this year
 

I'm guessing it may have come from the fact that the tangential velocity, with respect to the ground, of the top of a wheel, is twice the velocity of the vehicle to which it is connected.

So the reasoning goes: the roller must turn at 2x robot speed in order to keep the ball from escaping when backing up.

But that reasoning is flawed. The tangential velocity of the top of the wheel with respect to the robot is the same as the velocity of the robot.

There may be other reasons for the 2x number.


~

Re: The spread of pinching rollers this year
 

If you draw a ball and imagine it rolling at 1 m/s. At any moment in time, the bottom of the ball is stationary (in contact with the ground). The middle of the ball is traveling at 1 m/s. This means the top of the ball is at 2 m/s.

With a collector that touches the ball, it is the same. If the collector has the same surface speed as the robot is traveling, it will want to puch the ball forward as it draws it in. If the surfaces speed is 2x or greater, it will draw the ball into the robot. This is best seen with the backspin collectors when they are backing up. If the surface speed is 2 m/s, the robot will loose contact when its back-up speed is greater than 1 m/s. The ball still has some momentum and will continue towards the bot if the bot stops.

Normally the collectors suck the balls into some sort of feeder system. This year, there is a 3" intrusion rule. High speed collectors impart a lot of energy into the ball. This energy can cause a bounce when the ball hits its backstop. Some teams fixed this by moving from a top-spin to a lower point along the side of the ball. This reduced the rearward momentum, but induced higher normal forces between the ball and the carpet. This meant that you needed more power to keep the ball backspinning. The solutions for that were either a clutching mechanism (so that the ball did not backspin of which 1918 was extremely creative), more powerful motor, or larger ratio which meant slower surface speed. The topspin systems often used some sort of cushioned backstop to absorb the extra kinetic energy, or would have a cushioned roller mount (our original topspin, backspin system) with a lower idler roller.

With the high surface speed that 910 has, I imagine the had lots of testing with the ball rebounding out of the machine. It looks like they found a very interesting solution of compliant yet slick material. Again why that was my favorite collector this year.

If you need some diagrams, I can throw together another Powerpoint. I have a few decks of slides from our collector studies. I would have posted it up, but felt kind of silly saying, "Yeah, we did all this engineering to get this one design which was sort of OK, but then went with a design we ruled out in week 2 because we saw it was better in a video".

Re: The spread of pinching rollers this year
 

Carnegie Melon. Your design sounds like something that Carnegie Melon developed for the Robocup.